Sense checking the half-truth of attention metrics

There is no single definition for attention, which is a complex system with many interacting components, and there is a need to rethink how it is measured, says Bray’s Eaon Pritchard.

The Buridan’s Ass paradox is a thought experiment in philosophy that illustrates a situation involving a hypothetical donkey faced with two equal options. The paradox is named after Jean Buridan, a 14th-century French philosopher.

Imagine a donkey placed exactly halfway between two identical piles of hay. According to Buridan, if the donkey has to choose purely based on reason and logic, it should be unable to decide which option to take. The donkey would be caught in a perpetual state of indecision and ultimately starve to death.

This situation challenges the notion that a rational agent will always make a decision when faced with multiple equally desirable options.

Yet the donkey will not starve to death in the middle because its brain possesses other mechanisms that force it to select and behaviourally commit to its decision. According to this notion, attentional mechanisms are nature’s way of preventing organisms from being stuck in a Buridan’s Ass paradox.

Science or science-y?

“Data may well be the new oil but in the digital economy, attention is the scarcer resource.” This is the catchphrase of what many are calling the Attention Economy, and its measures, attention metrics, are the newest battleground in advertising.

It sounds impressive but what does it mean?

According to proponents of attention metrics, the access to media, especially the one we carry around in our pockets, and the almost total digital mediation of our media experience have created both the motive and the means for people to screen advertising out of their lives. And so traditional media currencies – things like impressions, reach and “opportunities to see” – are not that useful when the value of those metrics across channels and platforms is not the same.

Instead, these new attention measurements propose combining different data signals – things like duration of viewability, placement size, the surrounding page geometry and how many other ads are on the page. And the primary measurement tool for attention is eye-tracking software. 

The embedded eye-tracking tech follows and records what parts of the screen get looked at. Other signals – like for how long an ad is viewed, how much the viewer scrolled and if they clicked on the ad – can be layered onto the eye-tracking.

The theory is that when media planners use attention metrics, they can make better decisions on spending their advertising money based on how likely the placement will get the ad noticed.

But if the attention metrics lobby wants to claim this as an exact science, then we at least need an agreed definition of “attention”. But so far, we are only dealing in unreliable proxies.

In fact, there is no “one” definition. Attention is an incredibly complex system with many interacting components, some serving object perception, some deployed “volitionally” and some monitoring the environment in an ongoing manner for adaptively important situations. It’s the latter that should be of most interest to us.

The ecological view of attention (EVA) theory

The typical (yet erroneous) belief is that selective attention is a way to manage limited brain resources, that our senses collect more information than our brains can handle.

But paying selective attention requires a lot of brain power. Knowing this makes it seem odd to think of attention as a tool for freeing up brain resources. But on the contrary, limiting the processing of sensory information is beneficial, regardless of the brain’s capacity. This theory is the Ecological View of Attention (EVA); it focuses on how we interact with our surroundings rather than how much brain power we have.

The dominant force behind the evolution of selective attention is that without this mechanism, it would be difficult to distinguish targets from distractions. Attention mechanisms bypass this problem by ensuring that interference from distraction gets suppressed once a target is selected.

In that sense, the media environment is less important than what is being said if what’s being said matches your objectives.

The mind is complex, so a superordinate system is needed to coordinate the activity of neural systems, snapping each into the correct configuration at the right time. Emotions are principally functional states that regulate attention and behaviours.

Emotions evolved in response to conditions, contingencies, situations or events that recurred during evolutionary history. Things like avoiding and escaping from predators, parenting, exchange of trade and favours, establishing rank and status, and predicting other peoples’ behaviour. These are just a few.

Emotions are super-rational adaptations, finely tuned to countering threats and recognising opportunities. In “simple” terms, an emotion is a program whose function is to direct perception; inference; learning; memory; goal choice; motivational priorities and attention.

Heath’s low involvement processing theory

How much attention is needed, anyway? Perhaps, not a lot. Robert Heath proposed the Low Involvement Processing theory in the early 2000s and it may be the right time to revive this idea.

His theory argues that much advertising works not by persuading or informing consumers but by creating implicit associations that play out over time.

Crucially, much of this happens even when consumers are not actively paying attention to the ads and can’t recall them later. But the “recall” does happen at the shelf, in a buying situation, nonetheless.

Around the same time, cognitive psychology was also starting to recognise the importance of implicit memory, which influences behaviour without conscious awareness. Researchers found that stimuli could be processed and have an effect even without conscious attention or recall. This, in particular, is what tickled Heath, who argues that advertising can still be effective even in this state of low involvement.

This is because the exposure can still get “encoded” into memory, even if it’s peripheral or subconscious. It’s no accident that Coca-Cola invests a sizeable wad in providing signage for every corner shop and cafe on every street. It’s a final little nudge when you walk through the door because you are thirsty. But there’s no attention required.

Key to understanding memory and attention is that when we experience something, our brains don't store a perfect replica of that event. Instead, they piece together various fragments like sensations, emotions and perceptions to form a coherent memory. This process involves interpretation and filtering, meaning that the memory is constructed (and reconstructed) rather than an exact duplicate of the experience.

Rethinking attention

• Attention is an umbrella term for the mind’s set of operations that select some portions of a scene rather than others for more processing inside the mind. Fact.
• Attention evolved because some categories of information in the environment were likely to be more critical than others for activities that contributed to evolutionary goals. Fact.
• So whatever the media environment, advertising content that appeals to fundamental human evolutionary goals related to survival, reproduction, social interaction and status is more likely to attract attention. Fact.

There’s some interesting work going on with the attention metrics project. However, we are still unconvinced that alone, they give us anything more than proximate explanations. 

It is not yet credible as an industry standard “currency” due to the mystery around some of its methodologies and lack of consistency and evidence in the metrics used.

To avoid being sold snake-oil, the old methods of sniffing out BS still apply. Be sceptical, especially if it seems too neat or too good to be true.

Consider the credibility of the person or organisation presenting the information. Do they have expertise in the area they're discussing? Someone could be a decent media planner but that doesn’t make them a neuroscientist.

Look for corroborating evidence from credible sources. Is the same information being reported elsewhere by independent sources? Does any evidence actually exist?

Vague or generalised claims are more likely to be false than specific, detailed assertions. Everyone has a right to their opinion but not their own facts. As ever, the more extraordinary the claim, the more extraordinary the evidence to support it must be.